The Research On Intelligent Generating Paper Based On Genetic Algorithm

With the development of computer based education , computer examination system as an important component of computer managed instruction gets more and more attention. The study of generating algorithm to generate paper intelligently is a significant topic of computer based education.The test paper generating is an optimized problem to multi-objective parameter with certain constraint. It is very difficult that the optimization is implemented by traditional method. The quality and efficiency of intelligent generating test paper is all determined by the test questions-database designs and get problems-terms algorithm. A great deal of articles from inside and outside analyzed, Genetic Algorithm is selected as the way to generate test paper. The genetic algorithm is a kind of searching method using probability which simulates the natural evolution. Its predominance lies in effective resolving complicate and non-linear problems, which are difficult for traditional searching methods. It has few limitations on the presumption of the solution space, and owns predominance in adaptability and parallelism.This thesis briefly summarizes the rationales of genetic algorithm Firstly. Then, Research on the cause of premature convergence in this algorithm. Analyses the familiar defendable step and measure its variety. To solve the problems of premature convergence and blind genetic operators, based on the traditional genetic algorithm, this paper make a algorithm which lie on divided swarm tactic in macrostructure and lie on gene store tactic in microstructure assisted by the new unit tactic. The new algorithm improves capacity of regeneration schema and paces the algorithm converge.By analyzing the process of generate text paper, it establishes a mathematic model which combine the new genetic algorithm with generating paper problem. Ultimately design a Intelligent generating paper system for test. The system can search for the best answer according to such restriction conditions as test question types, terms scalar, knowledge points, difficulty degree, distinguish, exposition, the latest time and answer time. A cording method named paragraphed integer cording is used. Due to the cording strategy, crossover and mutation operator are designed. The actual test indicates that the intelligent generating test paper based on improved genetic algorithm can satisfy the needs for actual examinations, the speed is faster enough and the quality of test paper is better.